• Index of Linguistic Diversity:
• Vitality Index of Tradional Ecological Knowledge
  

D i s c u s s i o n

Differences among the three index components

	*see IBCD-RICH for full size map

IBCD-RICH. The most straightforward of the three components, IBCD-RICH produces index values that support earlier work showing that most of the countries highest in species richness are also among the highest in language richness. Indeed, the IBCD-RICH results extend the scope of this finding, for while Harmon 1996 compared only endemic species and languages, IBCD-RICH covers both non-endemic and endemic species and languages, while adding religions and ethnic groups. Of the ten highestranked countries for IBCD-RICH (see Figure 2) all but one are within the top thirty countries in each of the subcomponents of IBCD-RICH. The lone exception is Nigeria, which ranks 64th in PD-RICH. In fact, most rank within the top 15 countries in each of the subcomponents (Table 4). This points to a rather strong consistency across both cultural and biological diversity indicators, at least at the highest levels of IBCD-RICH.

By virtue of their first and second rankings in IBCD-RICH, Indonesia and Papua New Guinea together constitute the world’s leading “core area” of BCD richness. Several biogeographical factors—the presence of a vast archipelago, the highly variable terrain of New Guinea and the major Indonesian islands, and the fact that Wallace’s Line bisects the area—along with the absence (until recently) of a strong colonial presence, which enabled small hunter–gatherer groups to persist here in numbers perhaps larger than anywhere else, probably combine to explain the diversification of biological species and human cultures. Brazil–Colombia–Peru make up a second core area of BCD richness, with Nigeria–Cameroon–Democratic Congo making a third. It is interesting to note that all three are ecoregions dominated by tropical rainforests – Wallacea/New Guinea, the Amazon basin and the Congo basin. The other countries highest in BCD richness—India, China, USA, Mexico, and Australia—are all subcontinental (in Australia’s case, continental) in size and therefore encompass a large variety of ecosystems along with an array of indigenous cultures that have adapted to them, the latter producing high numbers of languages, religions, and ethnic groups.

Not surprisingly, the countries ranking lowest in IBCD-RICH (Figure 3) are all small islands (many of them in the Pacific), except for Greenland, which is large but heavily glaciated, and Cape Verde, Gibraltar, San Marino, three small mainland countries. Obviously, countries small in area are at a deep disadvantage in a tally based on richness alone.

The strengths of IBCD-RICH are:

• It is straightforward and easily grasped, relying on a simple count of entities (languages, religions, ethnic groups, and species) that people can readily understand.
• It requires no regression analysis of the data, as do IBCD-AREA and IBCD-POP.
• It produces index results that allow people to draw valid conclusions about the status of a country’s BCD richness. For example, compared to the other top-10 IBCD-RICH countries, Colombia ranks lowest in all three cultural diversity subcomponents of IBCD-RICH: 23rd in LD-RICH, 21st in RD-RICH, and 28th in ED-RICH). Yet Colombia ranks 10th in IBCD-RICH overall. Even if we knew nothing else, we could use this information to deduce that Colombia must not only be rich in species, but exceptionally rich, in order for it to make 10th overall in IBCD-RICH. And in fact that is precisely the case, for Colombia is the most species-rich country in the world, ranking 1st in both MD-RICH and PD-RICH.

Its weaknesses are:

• It is biased against small countries, particularly small island countries, and may lead to the mistaken impression that their BCD is somehow less important than that of larger countries.
• By depending solely on the number of bird, mammal and plant species, the biodiversity subcomponents of IBCD-RICH are biased towards higher vertebrates and plants at the expense of all other species. This is because birds, mammals and plants are the only taxonomic groups which have been comprehensively surveyed. Insects make up by far the largest group in the animal kingdom, and it is probable that there is a very large number of unrecorded species, particularly in the tropics. Plants have been reasonably well inventoried, but being a much larger group than either birds or mammals (an entire kingdom rather than a class) it is likely that many more plant species await discovery by science, again, particularly in tropical countries. Also, all marine species are excluded because of poor data for many countries, which therefore omits a large proportion of the biodiversity of islands and countries with extensive coasts, especially ones with tropical reef ecosystems.

Overall, IBCD-RICH appears to be a valuable means of measuring BCD, provided that appropriate caveats are given in terms of the accuracy of the data on which it is based, and of how the results are interpreted.

	*see IBCD-AREA for full size map

IBCD-AREA. By compensating for a country’s area—that is, by statistically neutralizing size differences so that both small and large countries can be considered on an equal footing—IBCD-AREA helps correct for the inherent bias against small countries that exists within IBCD-RICH. Countries that would have no chance to rank in the upper echelons of IBCD-RICH do so within IBCD-AREA: countries such as Togo and Israel, for example, both of which exhibit high LD-AREA, MD-AREA, and PD-AREA values but which could never rank highly in LD-RICH, MD-RICH, and PD-RICH simply because of their small size (see Table 5). It is instructive to see such countries appear high in a global ranking because it reminds us that even small countries have unique contributions to make to the overall complement of global BCD.

The strengths of IBCD-AREA are:

• Area-adjustment.
• The data manipulations are based on a widely accepted theory.

Its weaknesses are:

• It requires regression analysis and thus is not as straightforward or easy to understand as IBCD-RICH.
• As noted above, the species-area relationship and the entire theory of island biogeography are well accepted in ecology; nonetheless, they remain within the realm of theory and as such are subject to continuing reinterpretation and critique. For example, the theory of island biogeography is considered by some scientists to be inapplicable to certain taxa in certain places. The point here is not to impugn the species-area relationship, but simply to remind us that calculations (such as IBCD-RICH) that are based upon it are subject to revisions of the underlying theory.
• If used alone, IBCD-AREA may leave the false impression that certain small countries have greater overall BCD than larger ones (e.g., Brunei ranking just ahead of India, Nepal ahead of Brazil, etc.). IBCD-AREA is designed to correct for the biases in IBCD-RICH, and therefore the two must be used in concert.

	*see IBCD-POP for full size map

IBCD-POP. IBCD-POP is an extension of the species-area methodology upon which IBCD-AREA is based. IBCD-POP is an attempt to get at the relationship between human population size and the generation and maintenance of BCD. What this relationship is, remains murky; nevertheless, IBCD-POP generates some interesting results (see Table 6).

The strengths of IBCD-POP are:

• It assigns high index values to BCD-rich countries with relatively low human populations (e.g., French Guiana, Suriname, Guyana, Gabon), which is an intuitively plausible result. If, by contrast, densely populated and relatively BCD-homogeneous countries (e.g., Burundi, Bahrain) had achieved a very high ranking under IBCDPOP, one might be led to suspect that application of the basic species-area formula to population is entirely invalid.
• The fact that IBCD-POP produces results that seem to complement IBCD-RICH and IBCD–AREA (for more on this, see below) suggests that the IBCD–POP methodology is worth further investigation.

Its weaknesses are:

• It requires regression analysis and this is not as straightforward or easy to understand as IBCD-RICH.
• Extending the species-area relationship to per capita relationships is expedient and plausible, but nonetheless may turn out, upon further analysis, to be an invalid extension of the species-area formula; or at least may require some adjustment of that formula. In other words, the IBCD-POP methodology is promising but not proven.

Comparison of the three components: correlations. IBCD-RICH offers the most basic analysis of the available data. This method has both advantages and disadvantages. Simplicity is its most obvious virtue. However, it does not distinguish between countries or territories which have a high BCD only because they have a large land area or population and those which possess high diversity regardless of their land area or population. This is a disadvantage because countries are not being compared on a likefor- like basis. Is it surprising to learn that 119 languages are spoken in Russia, but only 7 in Reunion? IBCD-RICH does not shed much light on this question. A least-squares statistical analysis shows that there is a strong correlation (R² >0.6) between countries’ CD-RICH and BD-RICH values.

IBCD-AREA and IBCD-POP offer two alternative perspectives. IBCD-AREA is a robust method for analyzing biodiversity because the relationship between species richness and area (which was used to derive the index values of each country) is based on established ecological theory and observations, namely, that the number of species increases as a function of land area. It is reasonable to assume that the same relationship would be true for cultural diversity indicators. Interestingly, no single country or territory is more diverse than the world as a whole, after taking land area into consideration, for any of the five indicators used in IBCD-AREA. The global diversity value is therefore equivalent to the maximum index value. There is a good correlation (R² > 0.57) between number of languages and area, ethnic groups and area, bird/mammal species and area, and plant species and area. By contrast, there is only a moderate correlation between religions and area, and a poor correlation between CD-AREA and BD-AREA.

IBCD-POP is also based on the species-area relationship. While the analogous richnesspopulation relationship might be intuitively apparent between a country or territory’s cultural diversity and its population size, it is not obvious between biological diversity and human population. However, in IBCD-POP, there is a good correlation not only between language and population and ethnic groups and population, but also between birds/mammals and population, and plants and population. By contrast, there is only a moderate correlation between religion and population and a poor correlation between CD-POP and BD-POP.

The fact that the correlation between CD-AREA and BD-AREA and between CD-POP and BD-POP is relatively weak (R² = 0.20) means that countries with high cultural diversity do not necessarily have high biological diversity, and vice versa, after adjusting for either their land area or population size.Where there is no adjustment made, as in the IBCD-RICH index, there is a high correlation. Tables 7 and 8 show the actual values.

Comparison of the three components: rankings. Table 9, which summarizes the rankings for all three components, provides another basis for comparing the results among them. Perhaps the most striking aspect of the comparison is how consistently high Papua New Guinea and Indonesia rank under all three variants. Papua New Guinea ranks 2nd in IBCD-RICH, 2nd in IBCD-AREA, and 1st in IBCD-POP, with Indonesia ranking 1st, 1st, and 4th, respectively. By any measure, these two countries are the world leaders in BCD. Cameroon and Colombia are not far behind, being the only other two countries to rank in the top 10 under all three variants. When IBCD-RICH, -AREA, and –POP are themselves averaged (column 8 of Table 9), Papua New Guinea emerges slightly ahead of Indonesia, and so can lay claim to the title of the world’s most bioculturally diverse country —at least by these measures.

top

The world’s “core regions” of BCD

The world’s four most bioculturally diverse countries—Papua New Guinea, Indonesia, Cameroon, and Colombia—rank in the top ten for all three components of the index. Papua New Guinea ranks 2nd in IBCD-RICH, 2nd in IBCD-AREA, and 1st in IBCDPOP, with Indonesia ranking 1st, 1st, and 4th, respectively. By any measure, these two countries are the world leaders in biocultural diversity. Cameroon and Colombia are not far behind, being the only other two countries to rank in the top 10 in all three components.

By combining the results of BCD-RICH, BCD-AREA, and BCD-POP, we identified three “core regions” of global biocultural diversity that include countries of various sizes and populations (Figure 7):

• The Amazon Basin, consisting of Brazil, Columbia, and Peru, which ranked highly in BCD-RICH; Ecuador, which ranked highly in BCD-AREA; and French Guiana, Suriname, and Guyana, which ranked highly in BCD-POP.

• Central Africa, consisting of Nigeria, Cameroon, and the Democratic Republic of Congo (BCD-RICH), Tanzania (BCD-AREA), and Gabon and Congo (BCD-POP).

• Indomalaysia/Melanesia, consisting of Papua New Guinea and Indonesia (BCDRICH), Malaysia and Brunei (BCD-AREA), and Solomon Islands (BCD-POP).

Note that these regions are derived cumulatively; that is, they are geographic clusters centered on countries that are high in “raw” BCD richness (as measured by IBCD-RICH) to which adjacent countries highly ranked in IBCD-AREA and IBCD-POP are added. The resulting core regions are intuitively plausible in that they identify biogeographic realms that most experts would also identify as being among the most important for BCD: Indomalaya, the Amazon Basin, and Central Africa. We believe this is strong evidence that the three components of the IBCD give a more usable and realistic picture of where the world’s BCD is located than would an index based on raw BCD richness alone.

top

Deepening the analysis: trend data

It bears keeping in mind that the IBCD is founded on basic richness data: more or less straightforward counts of the products of cultural and biological diversity. As we discussed earlier, all global indices are built up out of simple information such as this. However, there is no reason why the IBCD could not be extended by additional empirical analysis of these (and other) indicators of BCD. The most useful information to add would be time-series statistics on the numbers of speakers of each language, members of each ethnic group and religion, and population sizes for each species. This additional data would allow an analysis of the distribution of abundance and therefore a more accurate estimate of diversity. More importantly, it would allow trends to be tracked over time.

For example:

• Trends in language use could be gauged by analyzing changes in the number of mother-tongue speakers of various languages, or by measuring changes in intergenerational transmission of language over time.
• Trends in religious adherence and ethnic group composition could be tracked by garnering demographic information on individual religions and ethnic groups.
• Trends in species could be supplemented by comparing the size of populations of species over time, or, in the absence of population data, by comparing species threat status from successive editions of the IUCN Red Books.
• Empirical data for other indicators of BCD, such as traditional environmental knowledge, could be gleaned (or solicited).

To illustrate what is possible, we next discuss in some detail how trends in language use might be studied, and briefly recount a recent empirical study of change in traditional environmental knowledge.

Global trends in numbers of mother-tongue speakers. One possible trend component would be a set of time-series data showing changes in the number of mother-tongue speakers of various languages. There are several sources of global language data dating back to the 1920s. The first (1924) edition of Meillet and Cohen’s Les Langues du Monde gives numbers of speakers for some languages and language families; Tesnière’s Statistique des Langues de l’Europe (1928) gives precise numbers for over 100 European languages. The second (1952) edition of Meillet and Cohen provides updated and more numerous figures. Other sources of global language data from years past include Voegelin and Voegelin 1977, Perrot 1981, Comrie 1987, Ruhlen 1987, and Gunnemark 1991.

Today, there are at least two published global language datasets: that of the Linguasphere Observatory (published in Barrett et al. 2001, and on-line at www.linguasphere.org), and the quadrennially updated Ethnologue series, published by the Summer Institute of Linguistics (now called SIL International). Ethnologue is probably the most widely cited source for global language data. In addition, UNESCO is undertaking a World Language Survey, which is as yet incomplete.

To get time-series data on the number of mother-tongue speakers of individual languages, we are assembling a dataset that is (to date) based primarily on three recent editions of Ethnologue. In previous work, one of the authors constructed a global database of the number of mother-tongue speakers based on the 1992 edition of Ethnologue, to which has been added data (for the Americas and Europe only, thus far) from the 1988 and 2000 editions and data (for Europe only) from Tesnière 1928.

Hence, so far we have compiled three (sometimes four) data points for many European languages (i.e., 1928, 1988, 1992, and 2000) and three data points for the languages of the Americas (1988, 1992, and 2000). To the extent that these data are an accurate reflection of the number of speakers of these languages, they can point us toward some long-term trends in language vitality. However, as with all cultural diversity data, a number of cautions are in order:

• The figures reported in Tesnière 1928 may not be directly comparable with those reported in recent editions of Ethnologue, because of terminological ambiguities, changes in classification, or differences in counting techniques. On the other hand, it seems likely that for small, clearly defined languages, Tesnière’s figures would be comparable to later Ethnologue figures.
• Within the editions of Ethnologue, changes in the number of mother-tongue speakers between 1988 and 2000 are often the result of better data becoming available rather than actual changes in populations, so for numerous languages apparent increases or declines are statistical artifacts and not reflections of reality.

Index of continuity and index of ability. In a quantitative study of vitality and moribundity, Statistics Canada used 1996 census responses to calculate an “index of continuity” and an “index of ability” for the country’s native languages. The index of continuity measures language vitality by comparing the number of people who speak it at home with the number who learned it as their mother tongue of origin. In this index, a 1:1 ratio is scored at 100, and represents a perfect maintenance situation in which every mother-tongue speaker keeps the language as a home language. Any score lower than 100 indicates a decline in the strength of the language. The index of ability compares the number who report being able to speak the language (at a conversational level) with the number of mother-tongue speakers. Here, a score of over 100 indicates that an increment of people have learned it as a second language, and may suggest some degree of language revival (Norris 1998, 10). Table 10 shows the main results of the study. All the elements of a thorough moribundity index are here: the size of the speaking population, indices of continuity and rejuvenation, and the average age of the speakers. By combining the two indices and adjusting the result by judiciously weighting the other factors, one could derive a quantitative measure of a given language’s vitality or lack thereof. Doing this on a global scale would require every country to conduct a census as thorough as Canada’s (which nonetheless still suffers from incomplete enumeration of some First Nations reserves).

Trends in Australia’s Aboriginal languages. Drawing on a wide range of studies and precepts (including those described above in Norris 1998), McConvell and Thieberger (2001) put together a status report on the Aboriginal languages of Australia—arguably the most endangered body of indigenous languages in the world. This report is in many ways a model of its kind, especially in terms of its comprehensive treatment of the factors that produce moribundity (and vitality) in small languages as they struggle to co-exist with a large, sociopolitically dominant language. While their full methodology is too detailed to be discussed here, it will suffice to note that they make good use of census and other data to develop age-class analyses of particular Aboriginal languages. From these age-class data, they create an Endangerment Index for these languages, which is the percentage of speakers aged 0-19 divided by the percentage of speakers aged 20-39. The intent, of course, is to try to see whether there is a drop-off in speaker percentage among the youngest generation. Languages with an index value of greater than 1 are considered “strong”; those with values of less than 1, “endangered.” McConvell and Thieberger go on to discuss a number of caveats and qualifications, especially concerning languages that had been considered “endangered” by earlier analysts (using different methods of analysis) but which rated higher than 1 in the Endangerment Index. These caveats, for example, point to potential problems with how census data on languages are to be interpreted. A main lesson from their study, which we noted at the outset of this paper, is that close familiarity with the situation “on the ground” and at relevant local scales is necessary for an accurate picture of cultural diversity.

Quantitative measurement of TEK change. The anthropologist Stanford Zent notes that in the ethnobotanical literature of the past two decades “it is extremely rare to find works that incorporate a time dimension” into studies of changes in traditional environmental knowledge (TEK) or even to find empirical studies of TEK change . To redress this, Zent carried out a study among the Piaroa, an indigenous ethnic group of Venezuela, which combined four research methods: (1) an ethnobotanical plot survey, (2) structured interviews, (3) informant consensus analyses, and (4) linear regression analyses. It will be seen that this research strategy is broadly interdisciplinary, combining botany, anthropology, and statistics to meld quantitative and qualitative information on a specific group at a subnational scale. Although the best way to measure TEK change would be through comparative baseline data, for many groups (such as the Piaroa) there is no information on TEK from an earlier, pre-disruption historical phase with which to compare contemporary changes. Zent reasons that evidence of variability within cultural knowledge foretells change in that knowledge, and so proposes an indirect method of inferring TEK change: “chart the pattern of knowledge variability within the Piaroa community” and then “study the relationship between this variability and social factors that are relevant indicators of the current situation of culture change”. Using the research strategy outlined above, Zent was able to demonstrate a drop in plant-naming competence among younger Piaroa and relate that decline to several social factors. This kind of technique holds promise for fine-grained studies of not only TEK, but of changes in other kinds of cultural diversity indicators for which quantitative timeseries data are difficult or impossible to get.

Deepening the analysis: endemism

In previous BCD research, comparisons have shown a high degree of overlap between the countries richest in endemic languages and those richest in endemic species. That research was based on data available in 1992. Here, we have updated data on endemic languages from the 2000 edition of Ethnologue and on species from the latest global biodiversity assessment . Although the three IBCD components presented here do not make use of data on endemics, we have included this information in the data tables to use as a springboard for discussing endemism. As a beginning, in Table 11 we have recalculated the concordance of the top 25 countries in species and language endemism in order to provide a comparison of rankings with the earlier study. In general, both the rankings and the concordance between the two top-25 lists remain the same. This kind of analysis, greatly expanded so as to discuss the implications of endemism for a global reckoning of BCD, could be part of an expanded version of the IBCD. (For more, see the Appendix, especially its concluding paragraph.)

top

next >>